BSD500

Context

The goal of this work is to provide an empirical basis for research on image de-noising

Content

The dataset consists of 500 natural images. This is a modification of the original BSD500 dataset which provides an empirical basis for research on image segmentation and boundary detection.

Acknowledgements
This dataset was obtained and modified from The Berkeley Segmentation Dataset and Benchmark from Computer Vision Group (University of California Berkeley). For more details on the dataset refer dataset's home page and related publication. Work based on the dataset should cite:

@InProceedings{MartinFTM01,
  author = {D. Martin and C. Fowlkes and D. Tal and J. Malik},
  title = {A Database of Human Segmented Natural Images and its
           Application to Evaluating Segmentation Algorithms and
           Measuring Ecological Statistics},
  booktitle = {Proc. 8th Int'l Conf. Computer Vision},
  year = {2001},
  month = {July},
  volume = {2},
  pages = {416--423}
}